I purchased a Ruuvitag sensor for monitoring relative humidity level in a humidor. The readings did not quite make sense, so I calibrated it.
I happen to have access to professional grade humidity generators and reference equipment.
As the humidor is supposed to remain at around ~70 %rH, did the following calibration run. The equipment used was Geo Calibrations 2000 SP humidity generator, with a traceably calibrated Vaisala HMP9 relative humidity and temperature probe as a reference. The test was made at +24 °C temperature. The Ruuvitag was removed from the plastic housing before calibration and the stabilization time in the chamber with flowing air was ~10 minutes for each measurement.
The temperature measurement was in very good agreement with the reference (around +/- 0.1 °C).
The humidity calibration results are:
Setpoint Ruuvitag deviation
50 +0.4 %rH
60 -1.2 %rH
70 -4.2 %rH
80 -7.0 %rH
90 -10.0 %rH
The off-the sleeve calibration uncertainty for this ad-hoc calibration setup is some +/- 4%-units. However as the run was made in one direction only, the effect of hysteresis is not visible.
Based on the results it seems the linearity of the sensor is OK and the accuracy is at the expected level for a sensor that is not factory calibrated. Bosch does not specify the higher range separately for BME280 but electronic humidity sensors in general are more inaccurate in the high humidity. Ruuvi does not specify the humidity accuracy of the final product so generally speaking I guess the sensor can be considered working as expected.
However, if better accuracy was required, both offset and gain (first-order) correction would be needed; at least for this particular sensor. So, as a future development idea, I suggest adding both offset (bias) and gain adjustment to the calibration functionality, so that the accuracy can be improved over a wider range.
For the time being I will make the offset adjustment as +4.2 %rH to achieve the best accuracy at the 70 %rH setpoint. This will be sufficient for my needs as the humidor is supposed to remain at this humidity. Who knows I will repeat the calibration sometimes in the future to see the long-term stability of the BME280.
